calculator-go

An application that calculates the confidence score for a given piece of data

Summary of Data Confidence scoring algorithm

1. Establish the total weighting of all factors

   • Defined by policy where AnnotationType = value from 1-5

     Example: "tpm"=2

   • If no policy or factor not found in policy, default weight will always be 1

   • For this working example, assume the following annotation types and weights

     tpm=2
     
     tls=1
     
     pki=1
     

   • Resulting total weights for the above where all annotations are present == 4

2. Establish which annotations are satisfied

   • Treat IsSatisfied == true as a 1 and multiply by the annotation's weight
   • For this example, TPM and TLS will be satisfied

     tpm -- IsSatisfied (1) Weight (2) == 2
     
     tls -- IsSatisfied (1) Weight (1) == 1
     
     pki -- IsSatisfied (0) Weight (1) == 0
     

   • Resulting total weight for the satisfied annotations == 3

3. Divide satisfied weight score by total weight score

   • 3 / 4 = .75 (%75 confidence)

Steps to Run OPA as server in docker container

1. Execute the following command inside the root directory of the project to build docker image from Dockerfile

   docker build -t opa-server scripts/policies
   

2. Execute the following command to opa-server container

   docker run --publish 8181:8181 opa-server
   

Steps to fetch the weights from OPA

1. Create input.json with the following content

   {"input":{"class":"production"}}
   

2. To execute a curl request using input.json

   curl localhost:8181/v1/data/dcf_scoring -d @input.json -H 'Content-Type: application/json' | jq
   

3. The response should be similar to

   {
     "result": {
       "weights": [
         {
           "pki": 2,
           "tls": 2,
           "tpm": 1
         }
       ]
     }
   }